Introduction
Mannheimia haemolytica (formerly Pasteurella haemolytica Biotype A) is a ruminant pathogen traditionally associated with severe respiratory disease of domestic sheep, bighorn sheep, and cattle, as well as septicemia in lambs and mastitis in ewes (Caswell and Williams, Reference Caswell, Williams and Maxie2007; Dassanayake et al., Reference Dassanayake, Shanthalingam, Herndon, Lawrence, Frances Cassirer, Potter, Foreyt, Clinkenbeard and Srikumaran2009, Reference Dassanayake, Shanthalingam, Herndon, Subramaniam, Lawrence, Bavananthasivam, Cassirer, Haldorson, Foreyt, Rurangirwa, Knowles, Besser and Srikumaran2010; Singh et al., Reference Singh, Ritchey and Confer2011; Besser et al., Reference Besser, Cassirer, Highland, Wolff, Justice-Allen, Mansfield, Davis and Foreyt2013; Gelasakis et al., Reference Gelasakis, Mavrogianni, Petridis, Vasileiou and Fthenakis2015). Bovine respiratory disease (BRD) is the major cause of beef cattle morbidity, mortality, and reduced production and costing the US cattle industry approximately $1 billion per year, when drugs, labor costs, decreased production, and animal death losses are taken into account (Marshall and Levy, Reference Marshall and Levy2011). In healthy cattle, M. haemolytica is a natural inhabitant of the upper respiratory tract including the nasal passages, nasopharynx, and tonsils; paranasal sinuses are predominately sterile, and M. haemolytica was isolated from transtracheal fluids from 13.1% of healthy cattle (Frank and Briggs, Reference Frank and Briggs1992; Frank et al., Reference Frank, Briggs and Zehr1995; Murray et al., Reference Murray, O'Neill, Lee, McElroy, More, Monagle, Earley and Cassidy2017; Timsit et al., Reference Timsit, Hallewell, Booker, Tison, Amat and Alexander2017). M. haemolytica live within biofilms on the upper respiratory mucosa (Olson et al., Reference Olson, Ceri, Morck, Buret and Read2002; Boukahil and Czuprynski, Reference Boukahil and Czuprynski2015, Reference Boukahil and Czuprynski2016, Reference Boukahil and Czuprynski2018). Multiple surface adhesins, including several surface proteins, fimbriae, and the polysaccharide capsule, are responsible for adherence of M. haemolytica to the upper respiratory mucosa and colonization (Morck et al., Reference Morck, Watts, Acres and Costerton1988; Jaramillo et al., Reference Jaramillo, Diaz, Hernandez, Debray, Trigo, Mendoza and Zenteno2000; Lo, Reference Lo2001; Gioia et al., Reference Gioia, Qin, Jiang, Clinkenbeard, Lo, Liu, Fox, Yerrapragada, McLeod, McNeill, Hemphill, Sodergren, Wang, Muzny, Homsi, Weinstock and Highlander2006; Daigneault and Lo, Reference Daigneault and Lo2009; Kisiela and Czuprynski, Reference Kisiela and Czuprynski2009). Early association of severe respiratory disease in stockyards and shipping earned the disease such names as stockyards pneumonia, shipping fever, and transit fever, whereas the name shipping fever is commonly used today for BRD in stressed beef cattle (Carter, Reference Carter1967; Mosier et al., Reference Mosier, Confer and Panciera1989a). Stress caused by environmental changes, shipping, weaning, comingling, and viral infections cause the bacterium to proliferate, to release from biofilms on the upper respiratory surface, and to be inhaled into the lower respiratory tract. Recently, an in vitro model demonstrated dispersal of M. haemolytica from biofilms treated with stress-related substances, epinephrine and to a lesser extent norepinephrine and substance P (Pillai et al., Reference Pillai, Cha and Mosier2018). When host defenses are overcome, the bacterium can precipitate severe fibrinous bronchopneumonia and death (Grey and Thomson, Reference Grey and Thomson1971; Frank and Smith, Reference Frank and Smith1983; Frank et al., Reference Frank, Briggs and Gillette1987; Caswell and Williams, Reference Caswell, Williams and Maxie2007; Booker et al., Reference Booker, Abutarbush, Morley, Jim, Pittman, Schunicht, Perrett, Wildman, Fenton, Guichon and Janzen2008; Panciera and Confer, Reference Panciera and Confer2010; Singh et al., Reference Singh, Ritchey and Confer2011). There are several serotypes of M. haemolytica (see “The organism”, below), and in severe, often fatal, pneumonia in cattle, especially in weaned beef cattle, Serotype 1 (S1) is most commonly isolated from sick cattle or from lesions of pneumonia (Purdy et al., Reference Purdy, Raleigh, Collins, Watts and Straus1997a; Al-Ghamdi et al., Reference Al-Ghamdi, Ames, Baker, Walker, Chase, Frank and Maheswaran2000; Katsuda et al., Reference Katsuda, Kamiyama, Kohmoto, Kawashima, Tsunemitsu and Eguchi2008; Panciera and Confer, Reference Panciera and Confer2010; Klima et al., Reference Klima, Alexander, Hendrick and McAllister2014). Serotypes 2 and 6 are common causes of sheep pneumonia (Odendaal and Henton, Reference Odendaal and Henton1995). Serotype 6 is isolated from BRD cases approximately 20% of the time or less, whereas S2 is often isolated from the nasal passages in high concentration in healthy, non-stressed cattle, but infrequently causes bovine pneumonia (Frank and Smith, Reference Frank and Smith1983).
The organism
M. haemolytica is a Gram-negative, non-motile, non-spore-forming, facultative anaerobic, weakly hemolytic coccobacillus, and a member of the family Pasteurellaceae (Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007). Several virulence factors are produced by the bacterium, notably endotoxin and leukotoxin (LKT) (see section below, “Virulence factors and potential immunogens”). It primarily infects ruminant species causing pneumonia, septicemia, or mastitis (Singh et al., Reference Singh, Ritchey and Confer2011). Historically, hemolytic strains of bacteria were isolated from bovine pneumonia in the 1920s and studied under the name Bacillus boviseptica (Jones, Reference Jones1921; Jones and Little, Reference Jones and Little1921). The organism was later named Pasteurella haemolytica to distinguish it from non-hemolytic Pasteurella multocida (Newsome and Cross, Reference Newsome and Cross1932). Studies to distinguish P. haemolytica strains serologically led to the classification of the bacterium into 16 serotypes using hemagglutination assays or a rapid direct, plate agglutination test (Carter, Reference Carter1956; Biberstein et al., Reference Biberstein, Gills and Knight1960; Biberstein, Reference Biberstein1965, Reference Biberstein1978; Frank and Wessman, Reference Frank and Wessman1978). Smith divided P. haemolytica into biotypes A and T, whereas biotype A strains fermented arabinose, and biotype T strains fermented trehalose (Smith, Reference Smith1959). In 1995, serotype A17 was added resulting in 13 serotypes as biotype A and 4 as biotype T (Younan and Fodar, Reference Younan and Fodar1995).
Bingham et al. (Reference Bingham, Moore and Richards1990) noted that biotype A strains were related by DNA homology, and biotype T strains were related; however, biotype A and T strains had a little genetic relationship. Sneath and Stevens (Reference Sneath and Stevens1990) proposed the name Pasteurella trehalosi for biotype T strains, and Blackall et al. (Reference Blackall, Bojesen, Christensen and Bisgaard2007) later demonstrated genotypically and phenotypically that those bacteria were distinct from other Pasteurella spp. and transferred P. trehalosi to a new genus as Bibersteinia trehalosi.
In 1999, the genus Mannheimia was proposed, and through DNA–DNA hybridization and 16S rRNA gene sequencing, all but one of the A biotypes were designated M. haemolytica. Pasteurella haemolytica A11 was unique and assigned a separate species, Mannheimia glucosida (Angen et al., Reference Angen, Mutters, Caugant, Olsen and Bisgaard1999). Therefore, M. haemolytica consists of the previous P. haemolytica biotype A Serotypes 1, 2, 5–9, 12–14, 16 and 17 along with untypable strains (Katsuda et al., Reference Katsuda, Kamiyama, Kohmoto, Kawashima, Tsunemitsu and Eguchi2008). Although all M. haemolytica serotypes are derived from biotype A, the designation of M. haemolytica serotypes as A1, A2, etc., often continues, even though it is redundant to include the biotype designation. Therefore, designation as Serotype 1 (S1), Serotype 2 (S2), Serotype 6 (S6), etc. seems appropriate. Serotypes 1 and 6 are closely related genetically and immunologically (Morton et al., Reference Morton, Panciera, Fulton, Frank, Ewing, Homer and Confer1995; Confer et al., Reference Confer, Ayalew, Panciera, Montelongo and Wray2006; Crouch et al., Reference Crouch, LaFleur, Ramage, Reddick, Murray, Donachie and Francis2012; Klima et al., Reference Klima, Cook, Zaheer, Laing, Gannon, Xu, Rasmussen, Potter, Hendrick, Alexander and McAllister2016). Despite genetic similarities among serotypes, sequencing of specific genes demonstrated diversity among similar serotypes and among bovine and ovine isolates. Davies et al. (Davies et al., Reference Davies, Whittam and Selander2001; Davies and Lee, Reference Davies and Lee2004) found that diversity exists in the major outer membrane protein OmpA between bovine and ovine isolates, and LKT diversity exists among ovine strains of M. haemolytica. Lawrence et al. (Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010) demonstrated an overall dissimilarity of 12% among LKT genes from several isolates. Ayalew et al. (Reference Ayalew, Blackwood and Confer2006) demonstrated that major outer membrane lipoprotein PlpE was highly conserved among M. haemolytica S1 and S6 strains and highly diverse among S2 strains. Comparison of the transferrin-binding protein operons (tbpBA) of M. haemolytica S1 and S6, M. glucosida, and B. trehalosi revealed the existence of a common gene pool among the organisms. In addition, tbpBA alleles of bovine M. haemolytica S1 and S6 are closely related to ovine origin strains (Lee and Davies, Reference Lee and Davies2011). A multiplex polymerase chain reaction (PCR) test for separating Serotypes 1, 2, and 6 was recently reported (Klima et al., Reference Klima, Zaheer, Briggs and McAllister2017).
Virulence factors and potential immunogens
M. haemolytica produces virulence factors that promote lung colonization, stimulate the production of inflammatory mediators, enhance evasion of host defense mechanisms, and stimulate an M. haemolytica – specific immune response. If the innate immune response fails to curtail pulmonary colonization, stimulation of pro-inflammatory cytokines and bacterial evasion of host defenses can promote development of pneumonia (Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007; Srikumaran et al., Reference Srikumaran, Kelling and Ambagala2007; Singh et al., Reference Singh, Ritchey and Confer2011). Virulence factors consist of capsular polysaccharides (CPS), lipopolysaccharide (LPS), adhesins, outer membrane proteins, iron-binding proteins, secreted enzymes, endotoxin, and the ruminant-specific repeats-in-toxin (RTX), LKT (Table 1) (Confer, Reference Confer2009). Klima et al. (Reference Klima, Alexander, Hendrick and McAllister2014) used PCR to screen for six M. haemolytica virulence genes including LKT (lktC), a putative adhesin (adhs), outer-membrane lipoprotein Gs60 (gs60), O-sialoglycoprotease (gcp), transferrin-binding protein B (tbpB), and UDP-N-acetyl-D-glucosamine-2-epimerase (nmaA). Each gene was identified in all M. haemolytica S1 and S6 isolates from both healthy and sick cattle. Finally, the importance of the various virulence factors, including CPS, LPS, adhesins, outer membrane proteins, iron-binding proteins, secreted enzymes, endotoxin, and LKT, in pathogenesis makes them targets of the host immune response and, therefore, potential targets for vaccine development.
Immunogens and potential immunogens of Mannheimia haemolytica
Adhesins
Specific M. haemolytica adhesins include a 68 kDa glycoprotein, N-acetyl-D-glucosamine, that mediates adherence to tracheal epithelial cells and activates the oxidative burst of bovine neutrophils through a 165-kDa glycoprotein receptor (Jaramillo et al., Reference Jaramillo, Diaz, Hernandez, Debray, Trigo, Mendoza and Zenteno2000; De la Mora et al., Reference De la Mora, Trigo, Jaramillo, Garfias, Solorzano, Agundis, Pereyra, Lascurain, Zenteno and Suarez-Guemes2006; De la Mora et al., Reference De la Mora, Suarez-Guemes, Trigo, Gorocica, Solorzano, Slomianny, Agundis, Pereyra and Zenteno2007). Heat-modifiable outer membrane protein A (OmpA) mediates M. haemolytica binding to bronchial epithelial cells and binds fibronectin, whereas addition of anti-OmpA antibodies reduces biofilm formation in vitro (Lo and Sorensen, Reference Lo and Sorensen2007; Kisiela and Czuprynski, Reference Kisiela and Czuprynski2009; Boukahil and Czuprynski, Reference Boukahil and Czuprynski2015). In addition, the 30-kDa surface Lipoprotein 1 was identified as important for M. haemolytica adhesion to the bronchial epithelium (Kisiela and Czuprynski, Reference Kisiela and Czuprynski2009). M. haemolytica capsule has antiphagocytic properties and may function as an adhesin (Morck et al., Reference Morck, Watts, Acres and Costerton1988, Reference Morck, Olson, Acres, Daoust and Costerton1989; Chae et al., Reference Chae, Gentry, Confer and Anderson1990; Whiteley et al., Reference Whiteley, Maheswaran, Weiss and Ames1990). Several other M. haemolytica adhesin proteins have been studied. A collagen-binding autotransporter adhesin was identified, and anti-autotransporter antibodies were in sera from M. haemolytica-challenged cattle (Daigneault and Lo, Reference Daigneault and Lo2009). Large, rigid fimbriae with subunit proteins of approximately 35 kDa and type IV pili were identified, and these structures often serve as adhesins in numerous bacterial species (Potter et al., Reference Potter, Ready and Gilchrist1988; Morck et al., Reference Morck, Olson, Acres, Daoust and Costerton1989; Lawrence et al., Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010). Filamentous hemagglutinins (Fha) are major surface proteins associated with adhesion of various bacteria, especially Bordetella spp. (Scheller and Cotter, Reference Scheller and Cotter2015). Gioia et al. (Reference Gioia, Qin, Jiang, Clinkenbeard, Lo, Liu, Fox, Yerrapragada, McLeod, McNeill, Hemphill, Sodergren, Wang, Muzny, Homsi, Weinstock and Highlander2006) demonstrated sequence homology among genes that code for M. haemolytica FhaB, in three M. haemolytica strains (Lawrence et al., Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010). Using in silico identification and high throughput screening, M. haemolytica Fha was identified as a highly immunoreactive protein when screened with sera against serotypes 1, 2, and 6 (Klima et al., Reference Klima, Zaheer, Cook, Rasmussen, Alexander, Potter, Hendrick and McAllister2018). The Serotype 1- specific (SSA-1) antigen was suggested to function as an adhesin. The gene coding for SSA-1 is expressed in vivo during lung infection and present in several M. haemolytica serotypes as well as in S1. In addition, the protein is highly immunogenic (Gonzalez et al., Reference Gonzalez, Murtaugh and Maheswaran1991, Reference Gonzalez, Maheswaran and Murtaugh1995; Lawrence et al., Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010; Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b; Sathiamoorthy et al., Reference Sathiamoorthy, Hodgins, Shewen, Highlander and Lo2011).
Secreted enzymes
M. haemolytica secretes numerous enzymes into culture supernatants and, therefore, likely the respiratory lumens during infection. Several proteases were recently identified in the culture supernatant of M. haemolytica S2, and those were primarily cysteine proteases or metalloproteases (Ramirez Rico et al., Reference Ramirez Rico, Martinez-Castillo, Gonzalez-Ruiz, Luna-Castro and de la Garza2017). A specific, 100 kDa Zn-dependent metalloprotease was identified in the same study.
Neuraminidase (sialidase) is an extracellular protein associated with numerous bacterial species and hydrolyzes sialic acid residues from host mucosal sialoglycoproteins exposing underlying carbohydrate moieties used for bacterial adhesion (Moncla et al., Reference Moncla, Braham and Hillier1990; Lewis and Lewis, Reference Lewis and Lewis2012). Therefore, neuraminidase is not a bacterial adhesin itself but may enhance bacterial adhesion through modification of cell surfaces allowing adhesins to interact with the surface. M. haemolytica neuraminidase was demonstrated to be a large, approximately 160 kDa, extracellular, heat-labile enzyme produced by various serotypes, primarily during stationary growth phase (Frank and Tabatabai, Reference Frank and Tabatabai1981; Straus and Purdy, Reference Straus and Purdy1995; Straus et al., Reference Straus, Purdy, Loan, Briggs and Frank1998; Highlander, Reference Highlander2001). Neuraminidase is produced in vivo in M. haemolytica-infected cattle as evidenced by the rise in anti-neuraminidase antibodies during infection (Straus et al., Reference Straus, Purdy, Loan, Briggs and Frank1998).
O-Sialoglycoprotease (also referred to as O-sialoglycoprotein endopeptidase and glycoprotease) is a 35.2 kDa endopeptidase that hydrolyzes peptide bonds within glycoproteins with a marked specificity for sialylated glycoproteins (Otulakowski et al., Reference Otulakowski, Shewen, Udoh, Mellors and Wilkie1983; Abdullah et al., Reference Abdullah, Lo and Mellors1990, Reference Abdullah, Udoh, Shewen and Mellors1992; Lee et al., Reference Lee, Lo, Shewen and Mellors1994a, Reference Lee, Shewen, Cladman, Conlon, Mellors and Lo1994b; Mellors and Lo, Reference Mellors and Lo1995). Its enzyme activity was originally identified in M. haemolytica culture supernatants, the protein was demonstrated by proteomic analyses in supernatants, and the enzyme was isolated and activity characterized as a neutral metalloprotease (Otulakowski et al., Reference Otulakowski, Shewen, Udoh, Mellors and Wilkie1983; Abdullah et al., Reference Abdullah, Udoh, Shewen and Mellors1992; Ayalew et al., Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b). Subsequently, Lo et al. (Reference Lo, Watts, Gyroffy and Mellors1994) cloned and expressed the gene as a fusion protein. The sialoglycoprotease gene and glycoprotease activity were associated with numerous M. haemolytica serotypes, and homologs of the protein were detected in several Gram-negative bacteria; however, secretion in the form of O-sialoglycoprotease was restricted to M. haemolytica serotypes (Lee et al., Reference Lee, Lo, Shewen and Mellors1994a; Lawrence et al., Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010; Klima et al., Reference Klima, Alexander, Hendrick and McAllister2014). Vaccination of calves with recombinant sialoglycoprotease-fusion protein stimulated antibodies against the protein (Shewen et al., Reference Shewen, Lee, Perets, Hodgins, Baldwin and Lo2003). Antibodies to sialoglycoprotease were identified in sera of cattle challenged with live M. haemolytica (Lee et al., Reference Lee, Shewen, Cladman, Conlon, Mellors and Lo1994b). The role of sialoglycoprotease in respiratory pathogenesis is yet unknown. It was shown to cleave cell surface glycoproteins CD34 (found on hematopoietic progenitors and endothelium), CD43 (leukosialin a surface protein on leukocytes), CD44 (a hyaluronic receptor serving as a cell adhesin molecule), CD45 (the leukocyte common antigen associated with signal transduction) and platelet selectin (Sutherland et al., Reference Sutherland, Abdullah, Cyopick and Mellors1992a, Reference Sutherland, Marsh, Davidson, Baker, Keating and Mellors1992b; Norgard et al., Reference Norgard, Moore, Diaz, Stults, Ushiyama, McEver, Cummings and Varki1993; Mellors and Lo, Reference Mellors and Lo1995). Lawrence et al. (Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010) suggested that the enzyme might assist in colonization of the upper respiratory tract.
Proteases that cleave host immunoglobulins have been described in various bacteria, are often produced as a component of autotransporter proteins, and can be secreted into the surrounding milieu (Mistry and Stockley, Reference Mistry and Stockley2006). Lee and Shewen (Reference Lee and Shewen1996) demonstrated bovine IgG1 protease activity in M. haemolytica culture supernatants that hydrolyzed bovine IgG1 into 39, 12, and 7 kDa bands, had no effect on IgG2, and was inhibited by EDTA, indicating it was a metalloprotease. The authors suggested that sialoglycoprotease might be involved in this process; however, further study on that point has not been documented in the literature. Proteases that cleave IgG have been identified in Streptococcus pyogenes, Pseudomonas aeruginosa, Proteus mirabilis, and Staphylococcus aureus (Brezski and Jordan, Reference Brezski and Jordan2010; Rungelrath et al., Reference Rungelrath, Wohlsein, Siebert, Stott, Prenger-Berninghoff, von Pawel-Rammingen, Valentin-Weigand, Baums and Seele2017; Wang et al., Reference Wang, Shi, Lv, Hu, Wang, Wang, Qiao, Zhang, Lv, Kjellman, Jarnum, Winstedt, Zhang, Wen, Hao and Yuki2017). A search of the genome databases of 10 M. haemolytica failed to demonstrate a putative IgG protease in M. haemolytica; however, numerous endopeptidases are present that could potentially cleave IgG (Ayalew, unpublished data, 2017). In addition, Ayalew et al. (Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b) demonstrated by proteomic analyses a putative IgA protease in M. haemolytica culture supernatant. IgA proteases in other bacteria are components of autotransporter molecules, enhance bacterial invasion of mucosal surfaces, and assist in bacterial escape from host defenses (Mistry and Stockley, Reference Mistry and Stockley2006). In addition, bacterial IgA proteases are immunogenic and stimulate local and systemic antibodies in infected hosts (Morelli et al., Reference Morelli, del Valle, Lammel, Pohlner, Muller, Blake, Brooks, Meyer, Koumare, Brieske and Achtman1994; Kirkeby et al., Reference Kirkeby, Rasmussen, Reinholdt and Kilian2000; Kotelnikova et al., Reference Kotelnikova, Zinchenko, Vikhrov, Alliluev, Serova, Gordeeva, Zhigis, Zueva, Razgulyaeva, Melikhova, Nokel, Drozhzhina and Rumsh2016).
Leukotoxin
M. haemolytica LKT (originally called cytotoxin) has been the subject of much research since M. haemolytica – induced cytotoxic damage of bovine macrophages and neutrophils was described in vitro, and the toxin was identified (Benson et al., Reference Benson, Thomson and Valli1978; Baluyut et al., Reference Baluyut, Simonson, Bemrick and Maheswaran1981; Berggren et al., Reference Berggren, Baluyut, Simonson, Bemrick and Maheswaran1981; Shewen and Wilkie, Reference Shewen and Wilkie1982). The LKT is encoded by four genes in the toxin operon, lktC, lktA, lktB, and lktD (Lo et al., Reference Lo, Strathdee and Shewen1987). lktA codes for the structural toxin, and lktC is involved in activation, whereas products of lktB and lktD are associated with secretion (Highlander, Reference Highlander2001; Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007). For a detailed description of M. haemolytica LKT and its genetics and activities, the reader is referred to one of several review articles (Highlander, Reference Highlander2001; Jeyaseelan et al., Reference Jeyaseelan, Sreevatsan and Maheswaran2002; Zecchinon et al., Reference Zecchinon, Fett and Desmecht2005; Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007; Czuprynski, Reference Czuprynski2009; Singh et al., Reference Singh, Ritchey and Confer2011). Recently, LKT acyl transferase gene (artJ-lktC) was used in a multiplex PCR assay that identifies M. haemolytica, P. multocida, and Trueperella pyogenes in infected lungs (Zhang et al., Reference Zhang, Liu, Liu, Ma, Xu and Wang2017).
After the discovery of M. haemolytica LKT, it was characterized as a 104 kDa protein and a member of the RTX family of toxins, which includes the Escherichia coli α-hemolysin and numerous toxins from other Gram-negative bacteria (Frey, Reference Frey2011). The toxic component of LKT resides in the N-terminus, whereas the region stimulating LKT–neutralizing antibodies is localized to a 32 amino acid region near the C-terminus (Lainson et al., Reference Lainson, Murray, Davies and Donachie1996; Welch, Reference Welch2001). The toxin is specific for leukocytes from various ruminant species and not for other cell types or leukocytes from nonruminant animal species; therefore, the name was changed from cytotoxin to LKT (Confer et al., Reference Confer, Simons, Barrie and Clinkenbeard1990; Jeyaseelan et al., Reference Jeyaseelan, Sreevatsan and Maheswaran2002; Narayanan et al., Reference Narayanan, Nagaraja, Chengappa and Stewart2002). LKT is secreted into culture supernatants during logarithmic growth phase, binds to the LKT receptor β2 integrins, CD18, and induces dose-related changes in bovine leukocytes (Shewen and Wilkie, Reference Shewen and Wilkie1985; Ambagala et al., Reference Ambagala, Ambagala and Srikumaran1999; Li et al., Reference Li, Clinkenbeard and Ritchey1999; Odendaal and Ellis, Reference Odendaal and Ellis1999; Dassanayake et al., Reference Dassanayake, Maheswaran and Srikumaran2007; Tucci et al., Reference Tucci, Estevez, Becco, Cabrera-Cabrera, Grotiuz, Reolon and Marin2016). At high LKT concentrations, leukocytes undergo rapid osmotic swelling, membrane pore formation, and necrosis (Clinkenbeard et al., Reference Clinkenbeard, Mosier and Confer1989). At reduced doses, LKT can induce leukocyte apoptosis, activate leukocytes with release of proinflammatory cytokines and oxygen-derived free radicals, reduce mitogen-mediated lymphogenesis, and stimulate histamine release from mast cells (Majury and Shewen, Reference Majury and Shewen1991; Maheswaran et al., Reference Maheswaran, Weiss, Kannan, Townsend, Reddy, Whiteley and Srikumaran1992; Adusu et al., Reference Adusu, Conlon, Shewen and Black1994; Cudd et al., Reference Cudd, Ownby, Clarke, Sun and Clinkenbeard2001, Reference Cudd, Clarke and Clinkenbeard2003; Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007; Singh et al., Reference Singh, Ritchey and Confer2011). Localization of LKT by immunohistochemistry in the lung after challenge demonstrated LKT associated with necrotic leukocytes and cell debris within alveoli (Whiteley et al., Reference Whiteley, Maheswaran, Weiss and Ames1990). Challenge of cattle with M. haemolytica LKT deletion mutants resulted in less severe lesions than with parent strains (Tatum et al., Reference Tatum, Briggs, Sreevatsan, Zehr, Ling Hsuan, Whiteley, Ames and Maheswaran1998; Highlander et al., Reference Highlander, Fedorova, Dusek, Panciera, Alvarez and Rinehart2000). Therefore, due to its many pathologic effects on leukocytes, LKT is considered the most important virulence factor in M. haemolytica – induced pneumonia.
M. haemolytica LKT is responsible for hemolysis in vitro and is produced by all serotypes of the bacterium with the exception of the occasional LKT – deficient mutant strains that have been described (Shewen and Wilkie, Reference Shewen and Wilkie1983b; Murphy et al., Reference Murphy, Whitworth, Clinkenbeard and Clinkenbeard1995; Ayalew et al., Reference Ayalew, Confer, Hansen and Couger2017a). LKT exposure stimulates neutralizing antibodies against the C-terminus region of the molecule. Despite some genetic diversity among LKT molecules, LKT-neutralizing antibodies against one M. haemolytica serotype or isolate usually neutralize LKT from another M. haemolytica serotype or isolate or from B. trehalosi (Gentry et al., Reference Gentry, Confer and Panciera1985; Lainson et al., Reference Lainson, Murray, Davies and Donachie1996; Hodgins and Shewen, Reference Hodgins and Shewen1998; Davies and Baillie, Reference Davies and Baillie2003; Shewen and Wilkie, Reference Shewen and Wilkie1983a, Reference Shewen and Wilkie1983b). Recently, however, it was shown that well-characterized LKT neutralizing monoclonal antibody MM601 did not neutralize serotype 2 or B. trehalosi serotype 10 LKT, and likewise, the monoclonal antibody raised against B. trehalosi T10 did not neutralize serotype 1 and 2 LKT, suggesting differences in their epitopes (Murugananthan et al., Reference Murugananthan, Shanthalingam, Batra, Alahan and Srikumaran2018). LKT-neutralizing antibodies strongly correlated with resistance against experimental challenge (Gentry et al., Reference Gentry, Confer and Panciera1985; Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007).
Capsule and lipopolysaccharide
M. haemolytica serotypes produce serotype-specific CPS that are on the surface of the bacteria, particularly during logarithmic growth (Corstvet et al., Reference Corstvet, Gentry, Newman, Rummage and Confer1982; Rice et al., Reference Rice, Carrasco-Medina, Hodgins and Shewen2007). S1 CPS is a complex mannose-rich polymer made of a disaccharide repeat of N-acetylmannosaminuronic acid β1,4 linked with N-acetylmannosamine, is moderately immunogenic, and partially protects the bacterium from phagocytosis (Adlam et al., Reference Adlam, Knights, Mugridge, Lindon, Baker, Beesley, Spacey, Craig and Nagy1984; Czuprynski et al., Reference Czuprynski, Noel and Adlam1989, Reference Czuprynski, Noel and Adlam1991a; Chae et al., Reference Chae, Gentry, Confer and Anderson1990; Conlon and Shewen, Reference Conlon and Shewen1993; Tigges and Loan, Reference Tigges and Loan1993). Purified M. haemolytica S1 CPS did not stimulate the release of pro-inflammatory cytokines from monocytes and macrophages (Czuprynski et al., Reference Czuprynski, Noel and Adlam1991b). Deposition of CPS into the lungs of sheep resulted in edema and a mild neutrophilic infiltrate with CPS binding to surfactant, whereas in lungs from cattle experimentally challenged with live M. haemolytica, CPS localized in alveolar lumens and macrophages, but not within the alveolar wall (Brogden et al., Reference Brogden, Adlam, Lehmkuhl, Cutlip, Knights and Engen1989; Whiteley et al., Reference Whiteley, Maheswaran, Weiss and Ames1990). Vaccination of calves with purified CPS, with or without other M. haemolytica antigens, stimulated significant anti-CPS IgM and IgG antibodies in sera; however, 36% of calves experimentally vaccinated with CPS developed anaphylaxis, and after challenge, lung lesion scores did not significantly correlate with anti-CPS titers (Conlon and Shewen, Reference Conlon and Shewen1993). IgG1 and IgG2 anti-CPS antibodies were highest in calves vaccinated with CPS in oil adjuvant, whereas IgM anti-CPS antibodies were highest in calves vaccinated with CPS with aluminum hydroxide adjuvant (Tigges and Loan, Reference Tigges and Loan1993). In another study, antibodies to purified CPS and to a partially purified saline extract that contained capsule, LPS, and proteins were examined in sera from cattle vaccinated with live or killed M. haemolytica (Confer et al., Reference Confer, Simons, Panciera, Mort and Mosier1989). Correlations between high antibodies to CPS and low lesion scores were inconsistent among experiments, whereas important antibodies in the saline extract were likely proteins (Confer et al., Reference Confer, Simons, Panciera, Mort and Mosier1989; Srinand et al., Reference Srinand, Maheswaran, Ames, Werdin and Hsuan1996b).
M. haemolytica LPS has a classical endotoxic activity that stimulates pro-inflammatory mediator production and inflammation. Those result in modification and damage of endothelium, causing vascular leakage, enhancement and depression of leukocyte functions, and complexing with LKT, increasing LKT-receptor production and augmenting LKT activity (Rimsay et al., Reference Rimsay, Coyle-Dennis, Lauerman and Squire1981; Confer and Simons, Reference Confer and Simons1986; Paulsen et al., Reference Paulsen, Mosier, Clinkenbeard and Confer1989; Reference Paulsen, Confer, Clinkenbeard and Mosier1990, Reference Paulsen, Confer, Clinkenbeard and Mosier1995; Kumar et al., Reference Kumar, Breider, Corstvet and Maddux1991; Saban et al., Reference Saban, Broadstone, Haak-Frendscho, Skoyen, Fialkowski, Maheswaran, Bjorling and Czuprynski1997; Cutlip et al., Reference Cutlip, Brogden and Lehmkuhl1998; Hsuan et al., Reference Hsuan, Kannan, Jeyaseelan, Prakash, Malazdrewich, Abrahamsen, Sieck and Maheswaran1999; Li and Clinkenbeard, Reference Li and Clinkenbeard1999; Lafleur et al., Reference Lafleur, Malazdrewich, Jeyaseelan, Bleifield, Abrahamsen and Maheswaran2001; Leite et al., Reference Leite, Gyles, Atapattu, Maheswaran and Czuprynski2003; McClenahan et al., Reference McClenahan, Hellenbrand, Atapattu, Aulik, Carlton, Kapur and Czuprynski2008). Intratracheal inoculation of M. haemolytica resulted in LPS within the cytoplasm of neutrophils, alveolar macrophages, endothelial cells, and pulmonary intravascular macrophages as well as on epithelial cell surfaces, as determined by immunohistochemistry (Whiteley et al., Reference Whiteley, Maheswaran, Weiss and Ames1990). Therefore, LPS is an important virulence factor widely distributed throughout the M. haemolytica-infected lung. Using monoclonal antibodies, antigenic similarities were demonstrated in a carbohydrate moiety of LPS extracted from serotypes 1, 5, 6, 7, 8, and 12 (Durham et al., Reference Durham, Antone, Cunningham and Confer1988). Nuclear magnetic resonance spectroscopy revealed the O-chain polysaccharides of S1, S6, and S9 to be identical and the core oligosaccharides of S1, S6, S8, S9, and S12 are similar (Lacroix et al., Reference Lacroix, Duncan, Jenkins, Leitch, Perry and Richards1993). Davies et al. demonstrated distinct serological differences among LPS molecules extracted from different serotypes of M. haemolytica (Davies and Donachie, Reference Davies and Donachie1996; Davies et al., Reference Davies, Arkinsaw and Selander1997); however, M. haemolytica LPS is poorly immunogenic, and no correlation existed between anti-LPS antibodies and resistance against experimental challenge (Confer et al., Reference Confer, Panciera and Mosier1986b). This may be because bovine M. haemolytica isolates often do not elaborate an O-antigen, which is the most immunogenic component of LPS (Ali et al., Reference Ali, Davies, Parton, Coote and Gibbs1992). Alternatively, the core lipopolysaccharide of M. haemolytica LPS has been analyzed, and glycoconjugates of LPS core were shown to be immunogenic in rabbits and stimulated complement-mediated killing of M. haemolytica (St Michael et al., Reference St Michael, Cairns, Filion, Neelamegan, Lacelle and Cox2011a, Reference St Michael, Vinogradov and Cox2011b).
Outer membrane proteins (OMPs)
The outer membrane of Gram-negative bacteria is a complex structure that assists bacteria to adapt to environmental changes, regulate influx and efflux of nutrients, and coordinate signal transduction (Khalid et al., Reference Khalid, Bond, Carpenter and Sansom2008). M. haemolytica has a host of proteins in the outer membrane, and many OMPs share partial sequence homology with OMPs from other Gram-negative pathogens (Squire et al., Reference Squire, Smiley and Croskell1984; Confer, Reference Confer1993; Davies et al., Reference Davies, McCluskey, Gibbs, Coote, Freer and Parton1994; Ayalew et al., Reference Ayalew, Confer, Hartson and Shrestha2010). Squire et al. (Reference Squire, Smiley and Croskell1984) extracted outer and inner M. haemolytica membranes using detergents and identified two major OMPs that were 30 and 42 kDa. Using radioiodination, Morton et al. (Reference Morton, Simons and Confer1996) demonstrated eight surface-exposed M. haemolytica proteins. Pandher et al. (Reference Pandher, Murphy and Confer1999) identified 21 surface-exposed, immunogenic M. haemolytica OMPs using Western immunoblots on protease-treated and untreated bacteria. Ayalew et al. (Reference Ayalew, Confer, Hartson and Shrestha2010) demonstrated 55 potentially immunogenic M. haemolytica OMPs by immunoproteomic analyses with those proteins potentially involved in cell structure, transport mechanisms, general metabolism, translation or other unknown functions. M. haemolytica outer membrane is adaptable with the number and character of the OMP profile changing depending on growth conditions and media (Gatewood et al., Reference Gatewood, Fenwick and Chengappa1994). OMPs are not virulence factors per se with the exception of OmpA and SSA-1, which have adhesin properties, and transferrin binding proteins that procure iron from host transferrin (Potter et al., Reference Potter, Schryvers, Ogunnariwo, Hutchins, Lo and Watts1999; Kisiela and Czuprynski, Reference Kisiela and Czuprynski2009; Lawrence et al., Reference Lawrence, Kittichotirat, McDermott and Bumgarner2010). The importance of OMPs for this review is as potential immune targets for the production of opsonizing antibodies to M. haemolytica. Shewen and Wilkie (Reference Shewen and Wilkie1988) demonstrated that vaccine immunity to M. haemolytica required both LKT-neutralizing antibodies and opsonizing antibodies to surface antigens. Because antibody responses to CPS and LPS do not appear to correlate with protection against M. haemolytica challenge, surface proteins are the more likely targets for stimulating the production of opsonizing antibodies. Several OMPs were examined as potential targets for the development of opsonizing antibodies. In fact, correlations between antibodies against several different OMPs and resistance to experimental challenge have been documented; therefore, opsonizing antibodies directed against multiple OMPs are likely involved in immunity to M. haemolytica (Mosier et al., Reference Mosier, Simons, Confer, Panciera and Clinkenbeard1989b).
M. haemolytica OMPs that have been studied to some degree with respect to immunity include OmpA, SSA-1, Gs60, PlpE, TBPs, PlpF, OmpD15, and OmpP2. Three approaches have been taken: (1) determination if antibodies to a specific OMP are present in higher concentration in sera from cattle that recovered from BRD compared with sera from cattle that were never ill, (2) determination if high antibodies to a specific OMP in sera from M. haemolytica–vaccinated cattle correlated with low lung lesion scores after experimental challenge, and (3) immunization of cattle with purified or recombinant OMPs either by themselves or in conjunction with a M. haemolytica vaccine followed by experimental challenge.
One of the first characterized M. haemolytica OMPs was the approximately 104 kDa SSA-1 (Gonzalez-Rayos et al., Reference Gonzalez-Rayos, Lo, Shewen and Beveridge1986). The protein was identified in E. coli expressing plasmids from a M. haemolytica gene library using antibodies against M. haemolytica culture supernatant. Initial studies indicated that the protein was reasonably specific for M. haemolytica S1; however, later studies identified ssa1 was distributed among seven M. haemolytica serotypes, including S1, S2, and S6 (Gonzalez et al., Reference Gonzalez, Murtaugh and Maheswaran1991). Subsequent studies indicated that the genes encoding SSA-1 derived from M. haemolytica S1 or S2 were identical (Gonzalez et al., Reference Gonzalez, Maheswaran and Murtaugh1995). Localization of SSA-1 in the outer membrane was confirmed in studies of M. haemolytica outer membrane vesicles (Ayalew et al., Reference Ayalew, Confer, Shrestha, Wilson and Montelongo2013; Roier et al., Reference Roier, Fenninger, Leitner, Rechberger, Reidl and Schild2013). Recently, Kumar et al. (Reference Kumar, Dixit and Kumar2015) used ssa1 gene as part of a multiplex PCR for rapid detection of M. haemolytica from sheep lungs. We vaccinated mice and cattle with recombinant SSA-1 and demonstrated that it was highly immunogenic resulting in significant increases in antibodies by day 28 after vaccination (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b). Vaccination of mice with recombinant SSA-1 and LKT-PlpE chimeric protein (SAC89) resulted in increased antibody responses to the SSA-1 and to the chimeric protein. Klima et al. (Reference Klima, Zaheer, Cook, Rasmussen, Alexander, Potter, Hendrick and McAllister2018) demonstrated that using in silico identification and high throughput screening of antigenic proteins, SSA-1 was the most immunoreactive of the proteins screened with antisera to M. haemolytica serotypes 1, 2, and 6. Challenge of cattle vaccinated with SSA-1 has not been done to our knowledge.
The OmpA Family of outer membrane proteins is a group of genetically related, heat-modifiable, surface-exposed, porin proteins ranging from 30 to 35 kDa that is in the outer membrane of numerous Gram-negative bacteria (Confer and Ayalew, Reference Confer and Ayalew2013). Members of the OmpA family of proteins are potential vaccine candidates for several bacteria (Pore and Chakrabarti, Reference Pore and Chakrabarti2013; Dubey et al., Reference Dubey, Avadhani, Mutalik, Sivadasan, Maiti, Girisha, Venugopal, Mutoloki, Evensen, Karunasagar and Munang'andu2016; Zhang et al., Reference Zhang, Yang, Cao, Sun, Dai and Zhang2016). OmpA for many bacteria is among the most numerous OMPs in the outer membrane (Khalid et al., Reference Khalid, Bond, Carpenter and Sansom2008). M. haemolytica OmpA is an approximately 30 kDa, heat-modifiable, surface-exposed, highly immunogenic protein with porin activity, and as a member of the OmpA family of proteins, it shares partial homology with heat-modifiable OMPs from numerous bacteria (Khalid et al., Reference Khalid, Bond, Carpenter and Sansom2008; Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b; Confer and Ayalew, Reference Confer and Ayalew2013). As described above, the protein has adhesin properties and was recently identified as binding lactoferrin, therefore, M. haemolytica OmpA may assist removal of iron from lactoferrin (Kisiela and Czuprynski, Reference Kisiela and Czuprynski2009; Samaniego-Barron et al., Reference Samaniego-Barron, Luna-Castro, Pina-Vazquez, Suarez-Guemes and de la Garza2016). M. haemolytica OmpA, originally called PomA, was purified and partially characterized, and its gene was cloned, expressed, and sequenced (Mahasreshti et al., Reference Mahasreshti, Murphy, Wyckoff, Farmer, Hancock and Confer1997; Zeng et al., Reference Zeng, Pandher and Murphy1999). Vaccination of cattle with live M. haemolytica resulted in high antibodies against M. haemolytica OmpA, and adsorption studies demonstrated surface-exposed epitopes (Mahasreshti et al., Reference Mahasreshti, Murphy, Wyckoff, Farmer, Hancock and Confer1997). Surface exposure of OmpA was corroborated, and two different OmpA subclasses (OmpA1 and OmpA2) with epitopic differences were identified in bovine and ovine isolates, respectively (Hounsome et al., Reference Hounsome, Baillie, Noofeli, Riboldi-Tunnicliffe, Burchmore, Isaacs and Davies2011). Vaccination of cattle with recombinant M. haemolytica OmpA stimulated high antibody responses with the complement-mediated killing of the bacterium (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b).
Gs60 is a surface-exposed, 60 kDa outer membrane lipoprotein found in M. haemolytica culture supernatant (Moore et al., Reference Moore, Hodgins, Firth, Mcbey and Shewen2011; Ayalew et al., Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b). A fragment of the gene was cloned, characterized, and expressed, and antibodies directed against the expressed protein fragment correlated with resistance against experimental challenge (Weldon et al., Reference Weldon, Mosier, Simons, Craven and Confer1994). Subsequently, the entire gene was cloned, sequenced, and found in P. haemolytica A biotypes (now M. haemolytica) and not in T biotypes (now B. trehalosi) (Lo and Mellors, Reference Lo and Mellors1996). In vivo expression of the gs60 and lkt genes were demonstrated within pneumonic lungs of cattle experimentally challenged with M. haemolytica (Lo et al., Reference Lo, Sathiamoorthy and Shewen2006). Lkt gene expression increased between 6 and 12 h after challenge, whereas gs60 gene expression decreased (Sathiamoorthy et al., Reference Sathiamoorthy, Shewen, Hodgins and Lo2012). The Gs60 protein was identified as a component of the putative M. haemolytica secretome but was not identified in an outer membrane immunoproteomics study of one M. haemolytica strain (Lo and Mellors, Reference Lo and Mellors1996; Ayalew et al., Reference Ayalew, Confer, Hartson and Shrestha2010, Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b). Recombinant Gs60 as a potential vaccine component was studied by several approaches including expression in alfalfa as a potential component of an edible vaccine and incorporation of M. haemolytica culture supernatant into immunostimulatory complexes (ISCOMs) with recombinant bovine C3d (Lee et al., Reference Lee, Cornelisse, Ziauddin, Slack, Hodgins, Strommer, Shewen and Lo2008; Moore et al., Reference Moore, Hodgins, Firth, Mcbey and Shewen2011). Feeding of dried Gs60-transgenic alfalfa to rabbits resulted in seroconversion to Gs60 (Lee et al., Reference Lee, Cornelisse, Ziauddin, Slack, Hodgins, Strommer, Shewen and Lo2008). In vivo cattle studies have not been reported with the Gs60 transgenic alfalfa, and Gs60 was not demonstrated in the ISCOMs. Further, in vivo studies demonstrated anti-Gs60 antibodies in sera from calves vaccinated with M. haemolytica supernatant vaccines and from calves vaccinated with recombinant Gs60 and challenged (Orouji et al., Reference Orouji, Hodgins, Lo and Shewen2012). In those studies, there were strong correlations between the production of antibodies to LKT and to Gs60, whereas increased antibodies to Gs60 were beneficial in resistance against challenge when anti-LKT antibodies were low.
Several immunogenic, surface-exposed M. haemolytica OMPs in the 40–50 kDa range were identified originally by several techniques (Mosier et al., Reference Mosier, Simons, Confer, Panciera and Clinkenbeard1989b; Morton et al., Reference Morton, Simons and Confer1996; Pandher et al., Reference Pandher, Murphy and Confer1999). Of those, the 45-kDa surface-exposed, outer membrane lipoprotein PlpE was extensively studied in our and other laboratories. The plpE gene was isolated from a gene library, cloned, sequenced, and expressed (Pandher et al., Reference Pandher, Confer and Murphy1998). Adsorption of serum antibodies with recombinant PlpE reduced the serum-mediated complement-mediated killing of M. haemolytica. The protein was named PlpE, because it was the fifth lipoprotein associated with the outer membrane of then P. haemolytica, with PlpA-C being three contiguous 28–30 kDa lipoproteins and PlpD being a 31-kDa lipoprotein (Cooney and Lo, Reference Cooney and Lo1993; Murphy and Whitworth, Reference Murphy and Whitworth1993; Dabo et al., Reference Dabo, Confer, Styre and Murphy1994; Murphy et al., Reference Murphy, Whitworth, Confer, Gaskins, Pandher and Dabo1998; Nardini et al., Reference Nardini, Mellors and Lo1998). PlpE was identified by immunoproteomics of the outer membrane and is in culture supernatants of M. haemolytica propagated under various growth conditions (Ayalew et al., Reference Ayalew, Confer, Hartson and Shrestha2010, Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b). Vaccination of cattle with recombinant PlpE plus adjuvant resulted in a reduction in lesion scores of >40% compared with controls after Mannheimia haemolytica S1 or S6 challenge (Confer et al., Reference Confer, Ayalew, Panciera, Montelongo, Whitworth and Hammer2003, Reference Confer, Ayalew, Panciera, Montelongo and Wray2006). In those studies, when incorporated with commercial M. haemolytica vaccines, recombinant PlpE enhanced resistance against challenge above that of the commercial vaccine alone. The immunodominant and potentially protective epitopes in PlpE are in a region of eight imperfect repeats of a hexapeptide in the N-terminal region (Ayalew et al., Reference Ayalew, Confer and Blackwood2004). Antibodies against that region stimulated complement-mediated killing of M. haemolytica. The 8 hexapeptide repeats were identical between M. haemolytica S1 and S6 isolates, whereas in S2 isolates, the repeats ranged from 3 to 28 hexapeptides (Ayalew et al., Reference Ayalew, Blackwood and Confer2006). Vaccination of mice and cattle with chimeric proteins composed of the hexapeptide-repeats epitope of PlpE and the neutralizing epitope of LKT stimulated antibodies that bound the surface of M. haemolytica and neutralized LKT (Ayalew et al., Reference Ayalew, Confer, Payton, Garrels, Shrestha, Ingram, Montelongo and Taylor2008; Batra et al., Reference Batra, Shanthalingam, Donofrio and Srikumaran2016a). Cattle vaccinated with chimeric protein plus formalin-killed bacteria were highly resistant against experimental challenge (Ayalew et al., Reference Ayalew, Step, Montelongo and Confer2009; Confer et al., Reference Confer, Ayalew, Montelongo, Step, Wray, Hansen and Panciera2009a, Reference Confer, Ayalew, Step, Trojan and Montelongo2009b; Guzman-Brambila et al., Reference Guzman-Brambila, Quintero-Fabian, Gonzalez-Castillo, de Obeso-Fernandez del Valle, Flores-Samaniego, de la Mora, Rojas-Mayorquin and Ortuno-Sahagun2012).
Iron is essential for bacterial growth and production of virulence factors, and bacteria have acquired several strategies for iron uptake (Gentry et al., Reference Gentry, Confer, Weinberg and Homer1986; Sheldon et al., Reference Sheldon, Laakso and Heinrichs2016). Strategies include extraction of iron from hemoglobin and acquisition from transferrin. In addition, through secretion and uptake of siderophores, free iron is obtained. M. haemolytica does not produce siderophores; therefore, M. haemolytica iron acquisition must be from heme, transferrin, and/or lactoferrin. During low iron concentrations, M. haemolytica produces iron-regulated OMPs (IROMPs). Thus, OMP profiles are different among bacteria grown in vitro in growth media that is iron-sufficient, iron-deficient, or iron-sufficient with an iron chelator. In addition, growth of bacteria in vivo, which is an iron-deficient environment, produces an OMP profile similar to that of the bacterium grown in vitro under iron-deficient conditions (Deneer and Potter, Reference Deneer and Potter1989; Morck et al., Reference Morck, Ellis, Domingue, Olson and Costerton1991; Confer et al., Reference Confer, Durham and Clarke1992, Reference Confer, McCraw, Durham, Morton and Panciera1995; Davies et al., Reference Davies, McCluskey, Gibbs, Coote, Freer and Parton1994; Gatewood et al., Reference Gatewood, Fenwick and Chengappa1994). LKT causes hemolysis of bovine erythrocytes, and in vivo transcription of two potential hemoglobin receptors, hmbR1 and hmbR2 were demonstrated in M. haemolytica within the lung (Murphy et al., Reference Murphy, Whitworth, Clinkenbeard and Clinkenbeard1995; Roehrig et al., Reference Roehrig, Tran, Spehr, Gunkel, Selzer and Ullrich2007). Iron acquisition from transferrin is a major mechanism used by M. haemolytica, and three IROMPs (approximately 70, 77, and 105 kDa) involved in transferrin binding were identified in bacteria grown in vitro under iron-restricted conditions or in vivo within an intraperitoneal implanted chamber (Deneer and Potter, Reference Deneer and Potter1989; Ogunnariwo and Schryvers, Reference Ogunnariwo and Schryvers1990; Morck et al., Reference Morck, Ellis, Domingue, Olson and Costerton1991; Yu et al., Reference Yu, Gray-Owen, Ogunnariwo and Schryvers1992; Geschwend et al., Reference Geschwend, Feist and Erler1997; Ogunnariwo et al., Reference Ogunnariwo, Woo, Lo, Gonzalez and Schryvers1997). Western blots using convalescent sera from M. haemolytica-infected calves demonstrated antibodies against the three proteins (Deneer and Potter, Reference Deneer and Potter1989; Puchalski et al., Reference Puchalski, Urban-Chmiel, Dec and Wernicki2013). We demonstrated antibody responses to the 70 and 77 kDa proteins to be significantly higher in live M. haemolytica-vaccinated calves than in control calves; however, there was no significant correlation between antibody responses to those proteins and lesion scores following challenge (Confer et al., Reference Confer, McCraw, Durham, Morton and Panciera1995). In another study, calves were vaccinated with either or both native 105 and/or recombinant 70 kDa proteins, termed transferrin-binding proteins (Tbp) A and B, respectively, and challenged (Potter et al., Reference Potter, Schryvers, Ogunnariwo, Hutchins, Lo and Watts1999). Both proteins were immunogenic and the best protection was in calves vaccinated with both TbpA and TbpB.
Through immunoproteomic analyses using 2D-electrophoresis and Western blots of M. haemolytica of outer membrane preparations probed with convalescent cattle sera, we identified several additional OMPs that were of interest for further study (Ayalew et al., Reference Ayalew, Confer, Hartson and Shrestha2010). These are PlpF, OmpD15, and OmpP2. PlpF is a 29.7 kDa lipoprotein that was identified in the first published M. haemolytica sequence as a conserved hypothetical protein (GI 7227128) (Highlander, Reference Highlander2001). The N terminus of PlpF contains a variable number of perfect and imperfect repeats, which varied among S1, S2 and S6 strains, and antigenicity plots predicted those repeats to be highly antigenic (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011a). The C-terminus half of PlpF shares substantial similarity with a surface-exposed, highly antigenic lipoprotein of Neisseria meningitidis (Madico et al., Reference Madico, Welsch, Lewis, McNaughton, Perlman, Costello, Ngampasutadol, Vogel, Granoff and Ram2006). Recombinant PlpF was immunogenic in mice and calves, and anti-PlpF antibodies are associated with complement-mediated killing (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011a). OmpD15 is an 88.8 kDa protein (also called Omp85 or Oma87) with homologues among various Gram-negative bacteria including Haemophilus ducreyi, Pasteurella multocida, H. influenzae, Shigella dysenteriae, Shigella flexneri, and Neisseria spp. (Ruffolo and Adler, Reference Ruffolo and Adler1996; Manning et al., Reference Manning, Reschke and Judd1998; Robb et al., Reference Robb, Orihuela, Ekkelenkamp and Niesel2001; Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b). Recombinant OmpD15 is immunogenic in mice; however, calves vaccinated with 100 µg in Freund's incomplete antigen developed only minimal antibody responses (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b). Unfortunately, challenge data from that study was lost due to a recording machine malfunction (Ayalew and Confer, unpublished data 2010). M. haemolytica OmpP2 (41.4 kDa) is a homologue of major outer membrane protein P2 of Haemophilus influenzae, which is known to have antigenic variation among H. influenzae isolates (Forbes et al., Reference Forbes, Bruce, Ball and Pennington1992; Andersen et al., Reference Andersen, Maier, Kemmer, Blass, Hilpert, Benz and Reidl2003). Both mice and calves vaccinated with recombinant OmpP2 developed low antibody responses suggesting that it may not be highly immunogenic at the dose or in the form that was administered (Ayalew et al., Reference Ayalew, Shrestha, Montelongo, Wilson and Confer2011b).
M. haemolytica vaccines
Realistic goals of vaccination
As described above, M. haemolytica produces various virulence factors that promote lung colonization, stimulate the production of inflammatory mediators, and enhance evasion of host defense mechanisms, as well as numerous potential immunogens that can stimulate an M. haemolytica–specific immune response. In 1975, Thomson et al. (Reference Thomson, Chander, Savan and Fox1975) demonstrated that on day 1 after shipping cattle that remained healthy had lower numbers of M. haemolytica in the nasal cavity and higher concentrations of anti-M. haemolytica antibodies than did cattle that became sick. Therefore, vaccination of cattle with efficacious M. haemolytica vaccines prior to shipment could potentially reduce shipping fever pneumonia. The overall goals for M. haemolytica vaccines are to stimulate an efficacious immune response that would (1) neutralize LKT and kill Mannheimia haemolytica, (2) reduce lung colonization, (3) block evasion of host defense mechanisms, and (4) reduce the severity of or prevent pneumonia. The central dogma of M. haemolytica vaccination was established by Shewen and Wilkie (Reference Shewen and Wilkie1988), whereas they demonstrated that efficacious vaccines must stimulate antibodies against LKT and against surface antigens, although the important surface antigens were not established in that study. Therefore, vaccines against M. haemolytica can accomplish this through multiple antibody-mediated mechanisms that (1) neutralize LKT, (2) enhance opsonization and phagocytosis, (3) block adhesion to respiratory cells, and/or (4) augment complement-mediated bacterial killing (Metzger, Reference Metzger, Kaufmann, Rouse and Sacks2011). It would be an added benefit if M. haemolytica vaccines also stimulated protection against multiple serotypes. Cross-serotypic LKT neutralization has been documented (Shewen and Wilkie, Reference Shewen and Wilkie1983b; Gentry et al., Reference Gentry, Confer and Holland1988). Surface antigens vary among serotypes making cross-serotypic protection often incomplete, but some degree of cross protection between other serotypes, especially for S1 and S6, has been demonstrated (Gentry et al., Reference Gentry, Confer and Holland1988; Morton et al., Reference Morton, Panciera, Fulton, Frank, Ewing, Homer and Confer1995; Purdy et al., Reference Purdy, Cooley and Straus1998; Confer et al., Reference Confer, Ayalew, Panciera, Montelongo and Wray2006; Crouch et al., Reference Crouch, LaFleur, Ramage, Reddick, Murray, Donachie and Francis2012; Zheng et al., Reference Zheng, Gupta, McCarthy, Moffat and Buddle2015). In addition, an added potential benefit of efficacious vaccination against bacteria is reduced use of antimicrobials and decreased development of antimicrobial resistance (Jansen et al., Reference Jansen, Knirsch and Anderson2018).
Due to the complexity of BRD, even with efficacy demonstrated in experimental challenge studies, determination of M. haemolytica vaccine efficacy can be a daunting task (Table 2). Rice et al. (Reference Rice, Carrasco-Medina, Hodgins and Shewen2007) described the difficulties in evaluation of vaccine efficacy field trials in beef calves, and especially the inadequacy of a single field trial to assess vaccine efficacy. For example, review of published vaccine field trials indicated that in some studies, vaccination of cattle with M. haemolytica bacterins enhanced protection against shipping fever, whereas in another study, no such protection was noted (Palotay et al., Reference Palotay, Young, Lovelace and Newhall1963; Amstutz et al., Reference Amstutz, Hortsman and Morter1981; Martin, Reference Martin1983; Perino and Hunsaker, Reference Perino and Hunsaker1997; Larson and Step, Reference Larson and Step2012). Various experimental challenge methods have been used to assess M. haemolytica vaccine efficacy. Originally, licensure of P. multocida and P. haemolytica bacterins used intraperitoneal vaccination and challenge in mice, which was not a good vaccine model for BRD and was discontinued (Confer, Reference Confer1993). Several experimental M. haemolytica challenge methods for cattle have been used, and the most common ones are aerosol, intratracheal, intrabronchial, or transthoracic routes with M. haemolytica alone or in combination with a respiratory virus such as bovine herpesvirus-1 (BHV-1), bovine viral diarrhea virus (BVDV), parainfluenza-3, or bovine respiratory syncytial virus (Frank, Reference Frank, Adlam and Rutter1989). Therefore, comparisons of different experimental vaccine efficacy trials can be confounded by variations in challenge methods. Additional considerations on vaccine trials are: (1) calves persistently infected with BVDV failed to mount an antibody response when vaccinated with a M. haemolytica bacterin-toxoid (Fulton et al., Reference Fulton, Step, Ridpath, Saliki, Confer, Johnson, Briggs, Hawley, Burge and Payton2003), and (2) M. haemolytica vaccines are often given with live or killed respiratory viral vaccines, and simultaneous vaccination of BHV-1 seronegative cattle with modified-live BHV-1 vaccine can interfere with the antibody response to M. haemolytica (Cortese et al., Reference Cortese, Seeger, Stokka, Hunsaker, Lardy, Weigel and Brumbaugh2011).
Various Mannheimia haemolytica vaccines tested under experimental and field trials
M. haemolytica bacterins
For an in-depth review of early vaccines directed against B. boviseptica, see Mosier et al. (Reference Mosier, Confer and Panciera1989a) and Mosier (Reference Mosier, Patten, Spencer, Johnson, Hoffmann and Lehane1992). In the early twentieth century, initial attempts to prevent BRD revolved around the use of bacterins, live vaccines, aggressins (bacteria-free inflammatory exudate produced by injection of virulent organisms), and antisera (Mosier et al., Reference Mosier, Confer and Panciera1989a). Several early reports indicated that antisera may reduce shipping fever mortality; however, in a later field study, use of prophylactic passive immunization after shipment had little protective value (King et al., Reference King, Edgington, Ferguson, Thomas, Pounden and Klosterman1955). Early vaccine studies often used small groups of cattle, and results suggested that both a bacterin and a live B. boviseptica vaccine could probably protect against natural disease; however, injection site abscesses were encountered with live B. boviseptica vaccines (Buckley and Gochenour, Reference Buckley and Gochenour1924). In one study, a 3-fold increase in death occurred in bacterin-vaccinated cattle compared with unvaccinated ones (Farley, Reference Farley1932).
Vaccination of cattle with M. haemolytica bacterins via various routes and with various adjuvants stimulated antibodies to the bacterium, and in several experimental studies, bacterins enhanced resistance against challenge, especially when oil adjuvants were used (Carter, Reference Carter1957; Wohler and Baugh, Reference Wohler and Baugh1980; Confer et al., Reference Confer, Panciera, Gentry and Fulton1987; Jericho et al., Reference Jericho, Cho and Kozub1990; Purdy et al., Reference Purdy, Straus and Ayers1997b). However, other studies indicated that M. haemolytica bacterin-vaccinated cattle were either not protected or had more severe disease than did unvaccinated controls when naturally or experimentally challenged (Hamdy and Trapp, Reference Hamdy and Trapp1964; Hamdy et al., Reference Hamdy, King and Trapp1965; Schipper and Kelling, Reference Schipper and Kelling1971; Friend et al., Reference Friend, Wilkie, Thomson and Barnum1977; Wilkie et al., Reference Wilkie, Markham and Shewen1980; Confer et al., Reference Confer, Panciera, Fulton, Gentry and Rummage1985b; Srinand et al., Reference Srinand, Ames, Maheswaran and King1995; Frank et al., Reference Frank, Briggs, Loan, Purdy and Zehr1996).
Live M. haemolytica vaccines
Interest in and study of live M. haemolytica vaccines developed in the 1980s. This was because of data showing inconclusive protection or enhanced disease in bacterin-vaccinated cattle, the discovery of LKT, which is not in bacterins, and demonstration that prior natural exposure of cattle to M. haemolytica enhanced resistance against challenge (Thomson et al., Reference Thomson, Chander, Savan and Fox1975; Shewen and Wilkie, Reference Shewen and Wilkie1982, Reference Shewen and Wilkie1983b; Confer et al., Reference Confer, Panciera and Fulton1984b). Live M. haemolytica vaccines studied contained wild type, attenuated, and chemically modified strains as well as a streptomycin-dependent mutant (Srinand et al., Reference Srinand, Ames, Maheswaran and King1995; Bowland and Shewen, Reference Bowland and Shewen2000). In several studies, parenteral or aerosol vaccination with live M. haemolytica resulted in enhanced resistance against experimental challenge. However, in field studies, live M. haemolytica vaccines either enhanced resistance or had no clear influence on morbidity and mortality (Confer et al., Reference Confer, Wright, Cummins, Panciera and Corstvet1983; Reference Confer, Panciera, Corstvet, Rummage and Fulton1984a, Reference Confer, Panciera, Fulton, Gentry and Rummage1985b, Reference Confer, Panciera, Gentry and Fulton1986a; Kucera et al., Reference Kucera, Wong and Feldner1983; Panciera et al., Reference Panciera, Corstvet, Confer and Gresham1984; Catt et al., Reference Catt, Chengappa, Kadel and Herren1985; Kadel et al., Reference Kadel, Chengappa and Herren1985; Smith et al., Reference Smith, Davidson and Henry1985; Purdy et al., Reference Purdy, Livingston, Frank, Cummins, Cole and Loan1986; Blanchard-Channell et al., Reference Blanchard-Channell, Ashfaq and Kadel1987; Srinand et al., Reference Srinand, Ames, Maheswaran and King1995, Reference Srinand, Maheswaran, Ames, Werdin and Hsuan1996b). Because LKT is produced primarily during logarithmic phase growth, one study demonstrated that in four of five experiments, aerosol vaccination with live M. haemolytica from 6-h cultures enhanced resistance to experimental challenge better than vaccination with 20–22 h live cultures (Confer et al., Reference Confer, Panciera, Corstvet, Rummage and Fulton1984a).
An intradermally administered attenuated strain, a chemically modified strain, and the streptomycin-dependent strain were commercialized M. haemolytica vaccines (Kucera et al., Reference Kucera, Wong and Feldner1983; Henry, Reference Henry1984; Panciera et al., Reference Panciera, Corstvet, Confer and Gresham1984; Confer et al., Reference Confer, Panciera, Corstvet, Rummage and Fulton1984a, Reference Confer, Panciera, Fulton, Gentry and Rummage1985b, Reference Confer, Panciera, Gentry and Fulton1986a; Smith et al., Reference Smith, Davidson and Henry1985). Use of the commercial intradermally administered attenuated M. haemolytica vaccine-enhanced resistance in experimentally challenged dairy calves and stimulated significant antibody responses to the bacterial surface in shipped beef calves. However, in one field trial, vaccination afforded marked protection, and in other field trials, the vaccination had no significant effect on performance, morbidity, or mortality (Confer et al., Reference Confer, Wright, Cummins, Panciera and Corstvet1983; Henry, Reference Henry1984; Smith et al., Reference Smith, Davidson and Henry1985; Purdy et al., Reference Purdy, Livingston, Frank, Cummins, Cole and Loan1986). The intradermal vaccine was later removed from the market partially because of difficulty in administration and local injection site reactions.
In one field trial, experimental live M. haemolytica vaccination by aerosol or subcutaneous routes prior to shipment stimulated significant antibody responses to the bacterial surface; however, significant protection against BRD was not demonstrated (Confer et al., Reference Confer, Wright, Cummins, Panciera and Corstvet1983).
Calves vaccinated with the chemically modified M. haemolytica strain had increased resistance to a BHV-1/M. haemolytica challenge when compared with nonvaccinated controls (Kucera et al., Reference Kucera, Wong and Feldner1983). Later, 11 cases of systemic M. haemolytica infection were described in post-vaccinated calves. Those calves had meningitis, polyarthritis, and dermatitis and/or cellulitis, and the infecting bacterial strain was identified by restriction endonuclease analysis as the chemically modified vaccine strain (Zeman et al., Reference Zeman, Neiger, Nietfield, Miskimins, Libal, Johnson, Janke, Gates and Forbes1993). That vaccine is no longer available.
Genetically modified M. haemolytica have been described including streptomycin-dependent, AroA deletion, and LKT – modified mutants. AroA is a component a metabolic pathway important in aromatic amino acid synthesis, construction of several AroA deletion mutants of M. haemolytica have been described (Homchampa et al., Reference Homchampa, Strugnell and Adler1994; Tatum et al., Reference Tatum, Briggs and Halling1994; Tatum and Briggs, Reference Tatum and Briggs2005). In one study, vaccination of mice with live M. haemolytica aroA− mutants reduced death in vaccinated mice; however, to our knowledge cattle studies have not been done (Homchampa et al., Reference Homchampa, Strugnell and Adler1994).
A bivalent vaccine containing streptomycin-dependent M. haemolytica and P. multocida mutants has been approved for parenteral or (recently) intranasal vaccination and marketed for many years (OncePMH®). (Catt et al., Reference Catt, Chengappa, Kadel and Herren1985; Kadel et al., Reference Kadel, Chengappa and Herren1985; Blanchard-Channell et al., Reference Blanchard-Channell, Ashfaq and Kadel1987; Chengappa et al., Reference Chengappa, McLaughlin, Kadel, Maddux and Greer1989; Mosier et al., Reference Mosier, Panciera, Rogers, Uhlich, Butine, Confer and Basaraba1998). Two studies demonstrated significant increases in antibodies to M. haemolytica following vaccination with the mutant bacteria as well as reduced clinical signs and lesions following BHV-1/M. haemolytica challenge (Catt et al., Reference Catt, Chengappa, Kadel and Herren1985; Blanchard-Channell et al., Reference Blanchard-Channell, Ashfaq and Kadel1987). Greater economic gains in vaccinated, non-preconditioned cattle were noted following a 50-day field trial (Kadel et al., Reference Kadel, Chengappa and Herren1985). Vaccination stimulated significant increases in antibodies to CPS and whole bacteria after one dose, whereas significant increases in antibodies against LKT and IROMPs required a booster vaccination (Srinand et al., Reference Srinand, Hsuan, Yoo, Maheswaran, Ames and Werdin1996a). In other studies, the streptomycin-dependent vaccine stimulated antibodies to M. haemolytica cell surface but not against LKT, and clinical and lesion scores for vaccinates were not significantly less than those in control cattle following transthoracic or intrabronchial M. haemolytica challenge (Srinand et al., Reference Srinand, Maheswaran, Ames, Werdin and Hsuan1996b; Mosier et al., Reference Mosier, Panciera, Rogers, Uhlich, Butine, Confer and Basaraba1998). Vaccinated veal calves had reduced respiratory morbidity compared with non-vaccinates (Schnepper et al., Reference Schnepper, Srinand and Jones1996).Vaccination of 14–20-day-old Holstein calves with the vaccine stimulated significant increases in antibodies; however, differences between vaccinated and control calves were not seen in BRD treatment data (Aubry et al., Reference Aubry, Warnick, Guard, Hill and Witt2001). Recently, a study compared cattle vaccinated with the streptomycin-dependent mutant vaccines by intranasal or subcutaneous routes and found no differences in cattle performance due to the route of vaccination (Spore et al., Reference Spore, Corrigan, Parks, Weibert, DeTray, Hollenbeck, Wahl and Blasi2017).
LKT-deficient M. haemolytica mutants have been studied as potential vaccines. Frank et al. demonstrated that intranasal exposure of shipped calves to live M. haemolytica with a 1-kb deletion in the lktA gene resulted in increased serum M. haemolytica antibodies and decreased M. haemolytica nasal colonization (Frank et al., Reference Frank, Briggs, Duff and Hurd2003). Recently, oral or parenteral vaccination of calves, sheep, and goats with live M. haemolytica mutants that produced N-terminal truncated LKT, which retained the neutralizing epitope, enhanced resistance against M. haemolytica challenge and stimulated both hemagglutinating and LKT-neutralizing antibodies (Briggs et al., Reference Briggs, Tabatabai and Tatum2012, Reference Briggs, Hauglund, Maheswaran and Tatum2013).
M. haemolytica extract vaccines
With the realization that M. haemolytica bacterins either failed to protect or caused enhanced disease, various antigen extraction procedures, including saline, potassium thiocyanate, and sodium salicylate, were studied to try to develop a better vaccine (Durham et al., Reference Durham, Confer, Mosier and Lessley1986).
Matsumoto et al. (Reference Matsumoto, Schmitz, Syuto, Watrous and Mattson1984) used a 2.5% saline extraction of M. haemolytica adsorbed to aluminum hydroxide gel as subcutaneous vaccine and extract alone for aerosol vaccination. They demonstrated enhanced resistance against a BHV-1/M. haemolytica challenge after subcutaneous vaccination; however, aerosol vaccination resulted in inconsistent results. Warm saline extraction of logarithmic-phase bacteria removed capsule and multiple surface proteins (Gentry et al., Reference Gentry, Corstvet and Panciera1982; Confer et al., Reference Confer, Lessley, Panciera, Fulton and Kreps1985a; Lessley et al., Reference Lessley, Confer, Mosier, Gentry, Durham and Rummage1985; McKinney et al., Reference McKinney, Confer, Rummage, Gentry and Durham1985). Vaccination with that saline extract enhanced resistance against transthoracic challenge, and there was a significant correlation between high antibodies to various protein components and low lesion scores (Confer et al., Reference Confer, Lessley, Panciera, Fulton and Kreps1985a; McKinney et al., Reference McKinney, Confer, Rummage, Gentry and Durham1985). Likewise, vaccination of cattle with a carbohydrate-protein subunit made by chromatofocusing of M. haemolytica saline extract enhanced resistance against transthoracic challenge (Confer et al., Reference Confer, Simons, Panciera, Mort and Mosier1989).
Sodium salicylate extraction of M. haemolytica S1 and S6 resulted in similar SDS-PAGE profiles as well as protein, carbohydrate, lipid, and phosphorus compositions (Donachie et al., Reference Donachie, Gilmour, Mould and Poxton1984). Vaccination with salicylate extracts with aluminum hydroxide adjuvant enhanced the resistance of calves and lambs against challenge with homologous serotypes (Gilmour et al., Reference Gilmour, Angus, Donachie and Fraser1982, Reference Gilmour, Martin, Sharp, Thompson, Wells and Donachie1983). In a later study, however, vaccination with salicylate extract failed to provide protection of calves against intranasal and intratracheal challenge and may have actually enhanced disease (Gilmour et al., Reference Gilmour, Gilmour, Donachie, Jones and Gourlay1987). Vaccination of calves with a salicylate extract containing IROMPs stimulated significant antibody responses to capsular polysaccharide and IROMPS, and those calves had significantly lower percent lung lesions than did controls after experimental challenge (Sreevatsan et al., Reference Sreevatsan, Ames, Werdin, Yoo and Maheswaran1996).
Potassium thiocyanate extracts of M. haemolytica as vaccines were studied briefly. Vaccination of mice and hamsters with M. haemolytica saline extract, potassium thiocyanate extract or bacterins indicated potassium thiocyanate extract resulted in greatest resistance against challenge (Tadayon and Lauerman, Reference Tadayon and Lauerman1981). Immunization of mice with a potassium thiocyanate extract of M. haemolytica enhanced cross-protection against P. multocida challenge (Mukkur, Reference Mukkur1977). Vaccination of calves via intranasal, subcutaneous, or intramuscular routes resulted in variable degrees of protection against aerosol BHV-1/M. haemolytica challenge (Yates et al., Reference Yates, Stockdale, Babiuk and Smith1983). Parenteral vaccination enhanced resistance and reduced bacterial isolation at necropsy better than did aerosol vaccination.
Capsular polysaccharide extract vaccines were tested experimentally. Vaccination of cattle with purified capsular polysaccharide stimulated antibodies to the capsule, and the intensity of the response and immunoglobulin type produced were dependent on the adjuvant used (Tigges and Loan, Reference Tigges and Loan1993). In another study, vaccination of calves with capsular polysaccharide alone or in conjunction with M. haemolytica culture supernatant or recombinant LKT did not protect calves (Conlon and Shewen, Reference Conlon and Shewen1993). In fact, capsular polysaccharide vaccination was associated with a 36% incidence of anaphylaxis. Others demonstrated that vaccination of calves with capsular polysaccharide with various dosages of muramyl dipeptide analogs stimulated resistance against challenge in several experiments; however, in an experiment comparing the capsular vaccine against commercial vaccines, the capsular vaccine had little efficacy (Brogden et al., Reference Brogden, DeBey, Audibert, Lehmkuhl and Chedid1995). More recently, authors suggested that (2→8)-α-Neu5Ac, which is a component of the capsule of Group B Neisseria meningitidis, E. coli K1, and M. haemolytica S2, might be used as a component of a conjugate vaccine (Robbins et al., Reference Robbins, Schneerson, Xie, Hanson and Miller2011).
A commercial vaccine that was a mild detergent extract of M. haemolytica was marketed for several years (Septimune® PH-K). The extraction method and detergent used were proprietary and not available in the literature. In published studies, the vaccine stimulated variable antibody responses to cell surface antigens and low anti-LKT antibodies (Confer and Panciera, Reference Confer and Panciera1994; Srinand et al., Reference Srinand, Hsuan, Yoo, Maheswaran, Ames and Werdin1996a; Confer et al., Reference Confer, Fulton, Clinkenbeard and Driskel1998). In two studies, Septimune enhanced resistance against M. haemolytica challenge, and in one study, vaccinated, shipped cattle had better, though not statistically significant, performance and health than did nonvaccinated cattle (Brogden et al., Reference Brogden, Adlam, Lehmkuhl, Cutlip, Knights and Engen1989; Hill et al., Reference Hill, Kirkpatrick, Gill and Ball1993; Confer and Panciera, Reference Confer and Panciera1994). That vaccine was later removed from the market. Saline, sodium salicylate, and potassium thiocyanate extracts likely lacked appreciable quantities of LKT to stimulate strong immunity and were not studied beyond the 1990s.
LKT supernatant vaccines
With the discovery of LKT secretion into culture supernatants and correlation between high LKT-neutralizing antibodies and resistance against the field or experimental BRD, Dr Bruce Wilkie's laboratory began to study the use of culture supernatant as a vaccine (Gentry et al., Reference Gentry, Confer and Panciera1985; Shewen and Wilkie, Reference Shewen and Wilkie1983a, Reference Shewen and Wilkie1983b). Besides LKT, M. haemolytica culture supernatant contains numerous secreted and surface antigenic proteins as well as capsular polysaccharide and LPS (Mosier et al., Reference Mosier, Simons, Chengappa and Confer1994; Mellors and Lo, Reference Mellors and Lo1995; Ayalew et al., Reference Ayalew, Confer, Hartson, Canaan, Payton and Couger2017b).
Shewen and Wilkie (Reference Shewen and Wilkie1988) demonstrated that vaccination with M. haemolytica S1 LKT-rich culture supernatant enhanced resistance against intrabronchial challenge with the homologous serotype, which led to the licensure of the commercial culture supernatant vaccine Presponse®. In that study, vaccination with LKT-rich culture supernatant derived from P. haemolytica S11 (now M. glucosida) stimulated LKT neutralizing antibodies but was not as efficacious against M. haemolytica S1 challenge. Those data led to the conclusion that antibodies to both LKT and surface antigens are important for enhancing resistance against M. haemolytica pneumonia. In addition, they noted that two doses of vaccine were more efficacious than one; however, in a 1995 study, one dose of the commercial vaccine was as efficacious as two doses against an intrabronchial challenge leading to licensure of Presponse for one-dose protection (Conlon et al., Reference Conlon, Gallo, Shewen and Adlam1995). The rationale for one-dose protection is based on a spontaneous rise in anti-M. haemolytica antibodies in young calves due to natural exposure through nasopharyngeal colonization; therefore, natural exposure is equivalent to primary vaccination (Hodgins and Shewen, Reference Hodgins and Shewen1998; Prado et al., Reference Prado, Prado, Payton and Confer2006). In another study, two doses of Presponse stimulated low peak antibody responses in 2-week-old, colostrum-deprived dairy calves; however, vaccinated calves had significant reductions in clinical signs and lesion scores compared with placebo-vaccinated calves (Hodgins and Shewen, Reference Hodgins and Shewen2000). In addition, Presponse vaccination with two doses of vaccine stimulated significant antibodies to surface antigens and to LKT, and vaccinated calves had a significant reduction in lung lesions after challenge when compared with controls (Sreevatsan et al., Reference Sreevatsan, Ames, Werdin, Yoo and Maheswaran1996). In the previously cited culture supernatant vaccine studies, anti-LKT and anti-surface antigen antibodies were demonstrated after vaccination. In contrast, Srinand et al. (Reference Srinand, Hsuan, Yoo, Maheswaran, Ames and Werdin1996a) demonstrated antibody responses to capsular polysaccharide but low to no anti-LKT or anti-whole cell antibodies in calves following Presponse vaccination. Similarly, we demonstrated that Presponse stimulated significant increases in antibodies to M. haemolytica whole cells and LKT as early as 7–14 days after vaccination; however, those responses were often lower than seen with other LKT-containing vaccines (Confer et al., Reference Confer, Fulton, Clinkenbeard and Driskel1998, Reference Confer, Montelongo, Brown, Fergen and Clement2001, Reference Confer, Ayalew, Panciera, Montelongo, Whitworth and Hammer2003).
As with many commercial vaccines, published studies of field trials are not numerous, and results can be variable due to conditions of the study, type of cattle used, and which respiratory pathogens may be involved in causing clinical disease. Bateman (Reference Bateman1988) vaccinated recently shipped, non-preconditioned calves with Presponse and found a slight decrease in morbidly, a slight improvement in treatment responses, and reduction in relapses. Jim et al. (Reference Jim, Guichon and Shaw1988) demonstrated reduced mortality, increased response to treatment, and economic benefits in feedlot calves vaccinated with Presponse In another study, Presponse vaccination of auction calves reduced relapse rates and mortality; however, vaccinated calves from ranches had no changes in morbidity rates or weight gains compared with nonvaccinated calves (Thorlakson et al., Reference Thorlakson, Martin and Peters1990). Average daily gains were improved in cattle vaccinated with Presponse at time of receiving compared with control cattle; however, morbidity and mortality data were not significantly different between vaccinated and nonvaccinated cattle (McLean et al., Reference McLean, Smith, Gill and Randolph1990). Brazle (Reference Brazle1992) vaccinated steers and bull calves and found no difference between vaccinated and controls with respect to weight gain, mortality, or morbidity. Fewer treatments, however, were required among vaccinates compared with controls. Malcolm-Callis et al. (Reference Malcolm-Callis, Galyean and Duff1994) demonstrated no benefit to vaccination in low morbidity calves but found increased gains and reduced treatments in vaccinated stressed calves. Bechtol and Jones (Reference Bechtol and Jones1996) suggested that Presponse vaccination of lightweight calves in a backgrounding lot was economically beneficial. Ives et al. (Reference Ives, Drouillard, Anderson, Stokka and Kuhl1999) found vaccination of calves with Presponse plus modified viral vaccines tended to reduce BRD incidence and retreatment rates, but those differences were not significant at P < 0.05.
In several experimental studies, Presponse was supplemented with either recombinant LKT, purified capsular polysaccharide, recombinant sialoglycoprotease, or recombinant outer membrane lipoprotein PlpE (Conlon et al., Reference Conlon, Shewen and Lo1991; Conlon and Shewen, Reference Conlon and Shewen1993; Confer et al., Reference Confer, Ayalew, Panciera, Montelongo, Whitworth and Hammer2003, Reference Confer, Ayalew, Panciera, Montelongo and Wray2006; Shewen et al., Reference Shewen, Lee, Perets, Hodgins, Baldwin and Lo2003). In those studies, the addition of any of the three recombinant proteins reduced clinical disease and/or lesion scores compared with Presponse alone, whereas addition of capsular polysaccharide failed to enhance resistance against challenge. In one experimental study, M. haemolytica culture supernatant was incorporated into polymerized methacrylic acid hydrogels and orally administered to calves (Bowersock et al., Reference Bowersock, Shalaby, Levy, Samuels, Lallone, White, Borie, Lehmeyer and Park1994). Vaccinated calves had less severe lung lesions and lived longer after intrabronchial challenge.
Bacterin-toxoid vaccines
Combinations of culture supernatants and killed M. haemolytica are marketed as bacterin-toxoid vaccines. Common examples include One Shot® and Pulmo-Guard®, and bacterin-toxoids are highly immunogenic stimulating intense antibody responses to surface antigens and LKT. One study demonstrated that One Shot vaccination also stimulated the production of the acute-phase protein haptoglobin (Arthington et al., Reference Arthington, Cooke, Maddock, Araujo, Moriel, Dilorenzo and Lamb2013; Moriel and Arthington, Reference Moriel and Arthington2013). Srinand et al. (Reference Srinand, Hsuan, Yoo, Maheswaran, Ames and Werdin1996a) demonstrated that a commercial bacterin-toxoid stimulated a significant antibody response to LKT, capsular polysaccharide, and whole cells but not to IROMPs. In numerous studies, high antibody responses to whole M. haemolytica and LKT were demonstrated after vaccination of cattle with One Shot or Pulmo-Guard (Loan and Tigges, Reference Loan and Tigges1989; Confer et al., Reference Confer, Fulton, Clinkenbeard and Driskel1998; Reference Confer, Montelongo, Brown, Fergen and Clement2001, Reference Confer, Ayalew, Panciera, Montelongo, Whitworth and Hammer2003, Reference Confer, Ayalew, Panciera, Montelongo and Wray2006; Mosier et al., Reference Mosier, Panciera, Rogers, Uhlich, Butine, Confer and Basaraba1998; Frank et al., Reference Frank, Briggs, Duff, Loan and Purdy2002; Fulton et al., Reference Fulton, Step, Ridpath, Saliki, Confer, Johnson, Briggs, Hawley, Burge and Payton2003; Ayalew et al., Reference Ayalew, Confer and Blackwood2004; Bowersock et al., Reference Bowersock, Sobecki, Terrill, Martinon, Meinert and Leyh2014).
Parenteral vaccination of cattle with M. haemolytica bacterin-toxoid vaccines enhanced resistance against experimental intrabronchial or transthoracic M. haemolytica S1 challenge and often-enhanced resistance more than other vaccines with which it was compared (Loan and Tigges, Reference Loan and Tigges1989; Confer and Panciera, Reference Confer and Panciera1994; Srinand et al., Reference Srinand, Hsuan, Yoo, Maheswaran, Ames and Werdin1996a; Mosier et al., Reference Mosier, Panciera, Rogers, Uhlich, Butine, Confer and Basaraba1998). Cattle vaccinated with One Shot, a bacterin-toxoid derived from M. haemolytica S1, had 46% reduction in lesion scores after challenge with M. haemolytica S6 when compared with control lesion scores (Confer et al., Reference Confer, Ayalew, Panciera, Montelongo and Wray2006). In that study, cattle vaccinated with One Shot supplemented with recombinant outer membrane lipoprotein PlpE had a 62% reduction in lesion scores compared with control cattle. Because of the relatedness between M. haemolytica and B. trehalosi. Bowersock et al. (Reference Bowersock, Sobecki, Terrill, Martinon, Meinert and Leyh2014) demonstrated that vaccination of calves with a multivalent modified-live virus vaccine containing One Shot enhanced resistance against an intrabronchial B. trehalosi challenge compared with vaccination with the virus vaccine alone.
There have been fewer published field trials with bacterin-toxoid vaccines compared with culture supernatant vaccine trials. Vaccination of cattle on entry to a feedlot with a H. somni – M. haemolytica bacterin-toxoid (SOMNU-STAR Ph, Elanco, Canada) resulted in increased antibodies to both bacteria and reduced morbidity (Van Donkersgoed et al., Reference Van Donkersgoed, Schumann, Harland, Potter and Janzen1993). Frank et al. (Reference Frank, Briggs, Duff, Loan and Purdy2002) found that M. haemolytica bacterin-toxoid vaccination prior to shipping did not enhance the resistance of cattle treated with florfenicol at an entry to the feedlot compared with nonvaccinated, florfenicol-treated calves. In field trials, Wildman et al. (Reference Wildman, Perrett, Abutarbush, Guichon, Pittman, Booker, Schunicht, Fenton and Jim2008) recommended the use of a bacterin-toxoid in conjunction with modified-live viral vaccines in a vaccination program for feedlot cattle. Vaccination of cattle with Pulmo-guard on arrival at a feedlot, significantly reduced mortality, but morbidity and average daily gain were unaffected by vaccination (MacGregor et al., Reference MacGregor, Smith, Perino and Hunsaker2003).
Autogenous vaccines
Several companies will make an M. haemolytica vaccine from a bacterial strain isolated from a specific farm, ranch, feedlot, or dairy. That vaccine is to be used only on those premises, and herd-specific vaccines have been used against several cattle pathogens (Attia et al., Reference Attia, Schmerold and Honel2013). Although these vaccines are not subject to the traditional safety and efficacy studies required for licensure of a commercial vaccine, they are required to be manufactured in a licensed facility and are subject to various regulations and guidelines for use that vary among countries (Attia et al., Reference Attia, Schmerold and Honel2013) (https://www.aphis.usda.gov/animal_health/vet_biologics/publications/pel_4_16.pdf). Published efficacy data on M. haemolytica autogenous vaccines are limited to an intraperitoneal M. haemolytica autogenous vaccine used effectively to control mastitis in sheep (Kabay and Ellis, Reference Kabay and Ellis1989). M. haemolytica autogenous vaccines are used in the field for BRD control; however, we and others have been unable to find published controlled studies of the use of M. haemolytica autogenous vaccines in BRD, and vaccine efficacies remain unknown (Miles and Rogers, Reference Miles and Rogers2014).
Experimental M. haemolytica vaccines
In recent years, several experimental approaches have been reported in attempts to develop improved M. haemolytica vaccines. These include outer membrane vesicles, recombinant proteins including chimeric proteins, and bacterial ghosts.
Gram-negative bacteria produce closed outer membrane blebs that detach as vesicles, which contain OMPs, LPS, periplasmic proteins, peptidoglycans, and secretory components such as toxins and have been studied as non-living, acellular vaccines against several bacteria (Kuehn and Kesty, Reference Kuehn and Kesty2005; Koeberling et al., Reference Koeberling, Seubert, Santos, Colaprico, Ugozzoli, Donnelly and Granoff2011; Nieves et al., Reference Nieves, Asakrah, Qazi, Brown, Kurtz, Aucoin, McLachlan, Roy and Morici2011; Park et al., Reference Park, Jang, Nho, Cha, Hikima, Ohtani, Aoki and Jung2011). M. haemolytica spontaneously produces vesicles in vitro, and proteomic analyses revealed 58 proteins of outer membrane or periplasmic membrane origin and LKT (Ayalew et al., Reference Ayalew, Confer, Shrestha, Wilson and Montelongo2013). Similarly, differences were minimal between M. haemolytica vesicle protein profiles and outer membrane protein profiles, and intranasal immunization of mice stimulated serum IgA and IgG1 antibody responses that reacted with SSA-1, OmpA, OMP P2, and several unidentified antigens (Roier et al., Reference Roier, Fenninger, Leitner, Rechberger, Reidl and Schild2013). Ramirez Rico et al. (Reference Ramirez Rico, Martinez-Castillo, Gonzalez-Ruiz, Luna-Castro and de la Garza2017) demonstrated that the culture supernatant of M. haemolytica S2 had higher protease activity than did outer membrane vesicles. Vaccination of calves with M. haemolytica vesicles stimulated antibodies to LKT and to surface antigens, and vaccinated calves had significant reductions in clinical signs and lesion scores after intrabronchial challenge (Ayalew et al., Reference Ayalew, Confer, Shrestha, Wilson and Montelongo2013).
As described in previous sections, addition of recombinant LKT, sialoglycoprotease, or PlpE to commercial vaccines enhanced resistance against experimental challenge compared with the vaccine alone (Conlon et al., Reference Conlon, Shewen and Lo1991; Confer et al., Reference Confer, Ayalew, Panciera, Montelongo, Whitworth and Hammer2003, Reference Confer, Ayalew, Panciera, Montelongo and Wray2006; Shewen et al., Reference Shewen, Lee, Perets, Hodgins, Baldwin and Lo2003). In each of those studies, vaccination of calves with only the recombinant protein demonstrated little or some beneficial effects, at least at the dosage given and with the adjuvant used. Feeding cattle with dried alfalfa expressing truncated LKT resulted in transient nasal IgA anti-LKT antibodies, and in a small pilot study, two orally vaccinated calves challenged with M. haemolytica had no lung lesions, whereas the two nonvaccinated controls had 11 and 27% pneumonic lesions (Shewen et al., Reference Shewen, Carrasco-Medina, McBey and Hodgins2009).
In several studies, recombinant chimeric or fusion proteins have been produced using the neutralizing epitope fragment of lktA and an immunogenic protein expressed from another gene. A recombinant protein with glutathione-S-transferase (GST), neutralizing epitope of LKT and Bordetella bronchiseptica fimbrial protein stimulated higher anti-LKT antibody responses in mice than did the GST-LKT protein minus fimbrial protein (Rajeev et al., Reference Rajeev, Kania, Nair, McPherson, Moore and Bemis2001). We produced several chimeric proteins from the neutralizing epitope of LKT and the major epitope of PlpE and demonstrated that mice vaccinated with those proteins developed antibodies against PlpE and LKT that had both complement-mediated bacterial killing and LKT neutralization activities (Ayalew et al., Reference Ayalew, Confer and Blackwood2004, Reference Ayalew, Confer, Payton, Garrels, Shrestha, Ingram, Montelongo and Taylor2008). Subcutaneous vaccination of cattle with a PlpE-LKT chimeric (SAC89) plus formalin- killed bacteria in an oil-in-water adjuvant resulted in 75% lower lesion scores compared with controls, whereas vaccination with the chimeric protein or bacterin alone resulted in approximately 35% lower lesion scores compared with controls (Confer et al., Reference Confer, Ayalew, Montelongo, Step, Wray, Hansen and Panciera2009a). Subsequently, intranasal vaccination of cattle with PlpE-LKT chimeric protein plus cholera toxin stimulated nasal anti-whole cell and anti-LKT antibodies, whereas intranasal vaccination of calves with a PlpE-LKT-cholera toxin subunit B chimeric protein (SAC102) stimulated serum and nasal antibodies (Ayalew et al., Reference Ayalew, Step, Montelongo and Confer2009; Confer et al., Reference Confer, Ayalew, Step, Trojan and Montelongo2009b). Calves vaccinated with SAC102 had lower clinical signs after intrabronchial challenge than did nonvaccinated calves. Recently, a similar PlpE-LKT chimeric protein stimulated anti-PlpE and anti-LKT antibodies in mice (Batra et al., Reference Batra, Shanthalingam, Donofrio and Srikumaran2016b). Those scientists intranasally vaccinated Bighorn sheep with a recombinant BHV-1-vectored vaccine expressing PlpE-LKT chimeric proteins. Vaccinated sheep developed anti-LKT antibodies, but inconsistently developed anti-surface antibodies, and the vaccine failed to protect against M. haemolytica challenge (Batra et al., Reference Batra, Shanthalingam, Donofrio, Haldorson, Chowdhury, White and Srikumaran2017).
Bacterial ghosts are a non-living vaccine strategy, wherein bacteria are infected with a temperature-controlled lytic phage that causes membrane tunnels through which the bacterial genome (cytosol?) is expelled leaving a bacterial envelop (ghost) that has proteins that were not modified by exposure to formalin or another bactericidal substance, as with a bacterin (Szostak et al., Reference Szostak, Hensel, Eko, Klein, Auer, Mader, Haslberger, Bunka, Wanner and Lubitz1996; Lubitz et al., Reference Lubitz, Witte, Eko, Kamal, Jechlinger, Brand, Marchart, Haidinger, Huter, Felnerova, Stralis-Alves, Lechleitner, Melzer, Szostak, Resch, Mader, Kuen, Mayr, Mayrhofer, Geretschlager, Haslberger and Hensel1999). Vaccination of mice and calves with M. haemolytica ghosts plus adjuvant enhanced resistance against challenge similar to a commercial vaccine (Pastobov®, Merial) (Marchart et al., Reference Marchart, Dropmann, Lechleitner, Schlapp, Wanner, Szostak and Lubitz2003a, Reference Marchart, Rehagen, Dropmann, Szostak, Alldinger, Lechleitner, Schlapp, Resch and Lubitz2003b).
Conclusions
In the last 30 years, much has been learned about immunogens and potential immunogens of M. haemolytica. Current commercial vaccines, in general, are improvements over prior bacterins, albeit field trials have not always demonstrated efficacy in the face of a complex of pathogens and environmental stressors. Enough data have been generated on immunogenic recombinant M. haemolytica proteins that warrant further studies to develop a new generation of M. haemolytica vaccines with increased efficacy beyond that experienced with today's vaccines.
